500 
What is so important to consider in the study of double 
decomposition is that the fact, that the sum of the masses of the 
two products of the change is the sum of the masses of the two 
interacting substances, presents itself no longer as being merely 
the evidence of the massing together of substances into a com- 
pound ; for there is in double decomposition to be considered 
that redistribution of mass which, on the one hand, is found to 
correspond to and be part of a general though not sharply 
defined redistribution of physical and chemical properties ; and, 
on the other hand, to be obviously irreducible to that inter- 
change of those simpler substances which in many cases are 
produced in the simple decomposition of the acting substances. 
The physical properties of substances, or rather their sensible 
qualities, are of too uncertain a character for their redistribution 
to be safely traced. But it generally does result, amongst 
inorganic substances, at least, that colour is transmitted, the 
saline, acid, bitter, or other taste of one of the active substances 
will appear, with more or less distinctness, in one of the pro- 
ducts, a relatively volatile and a relatively fixed substance 
together will yield a similar pair of products, a dense and a light 
substance will yield a dense and a light substance, and so on. 
The chemical properties, however, are quite definitely re- 
distributed to a large extent, a fact sufficiently illustrated by 
saying that an iron salt yields an iron salt, and a sulphate yields 
a sulphate. 
But this is not a redistribution in which simpler substances, 
or indeed any other substances than those interacting, play a 
part; as soon becomes evident on attempting to establish the 
contrary by an appeal to the facts. While silver acetate and 
silver sulphate resemble each other and also silver nitrate as 
silver salts, they do not resemble silver itself ; and though 
silver nitrate resembles sodium nitrate as nitrate, there is not 
even a substance known which is related to these salts as silver 
is related to silver salts. It might be objected to this that 
there may yet become known such a substance, which in its 
ultimate decomposition would give one molecule of nitrogen to 
three molecules of oxygen. If instead of nitrate were given 
acetate or cyanide, there would be found in the substances 
acetic peroxide and cyanogen, it might be said, the analogues 
of the as yet unknown substance of the nitrate. But the point 
I would make is that nitrate, sulphate, &c., are names with 
well-defined meanings independent of the fact that the corre- 
sponding substances are not known; for it follows without 
argument that also the terms silver, iron, chloride, &c., should 
be equally independent in meaning of the existence of the sub- 
stances silver, sodium, chlorine, &c. It is a familiar historical 
fact that cesium, helium and fluorine were chemical names 
long before the substances cesium, helium and fluorine became 
known. We might well be convinced, therefore, without going 
further, that constitutional names, names which convey the 
facts of likenesses preserved in chemical change, cannot be 
indicators of the presence of the substances for which they may 
be also used. For, that being the case, we have no grounds 
for assuming that silver nitrate in interaction with sodium sul- 
phate decomposes into the substance silver, which then com- 
bines to form silver sulphate. But fuller proof than any 
appearance of likeness or unlikeness can give is afforded by 
facts which became known and appreciated in connection with 
the chemical molecule. Typical of them all is the fact that in 
none of its interactions does chlormethane yield a hydrocarbon 
simpler than methane or than itself. Under those conditions in 
which it might have been expected to give a substance which 
would be methyl, it produces ethane, a substance which chlorine 
converts into another substance, having instead of one-third 
only one-sixth less hydrogen in its composition. Similar results 
have been obtained in all cases where the point can be deter- 
mined—that is, where the simpler substance looked for would 
still be a compound substance, and such simpler derivatives are 
looked for no longer. The monohydride of oxygen or sulphur, 
the dihydride of nitrogen or phosphorus or arsenic, the mono- | 
nitride of carbon, the organic compounds, methyl, phenyl, 
acetyl, are not only unknown, but are held to be non-existent 
substances, though their chemical compounds, the hydroxides, 
amides, cyanides and the rest are both numerous and well 
defined. , Whatever other view we shall have to take of the 
constitution of Gomberg’s remarkable ‘‘ triphenylmethyl,” it 
will certainly not be that it is identical with the radical of the 
triphenylchlormethane from which it is derived, unless we are 
prepared to allow that carbon is sometimes tervalent. Ethylene 
the substance differs from ethylene the radical in having its two 
NO. 1716, VOL. 66] 
NATURE 
[SEPTEMBER 18, 1902 
carbons differently related ; butit is difficult to see how to make 
a similar distinction in the case of Gomberg’s substance. 
In those other cases in which the point is not strictly deter- 
minable, only because the resulting substances are the simple 
substances themselves, it required but the recognition of 
molecular quantities to make it evident that these cases run 
parallel with the others. For, in all changes which can be 
satisfactorily followed out, the resulting or entering quantity of 
the simple substance is twice as great as that which can have 
come from, or gone to form the molecule of either of the 
compound substances. But if, so far as can be traced, a simple 
substance comes only half from one molecule of any of its 
compounds, none of these compounds can contain or be com- 
posed of simple substances. All simple substances, therefore, 
as well as all compound substances, enter into and come out 
from chemical changes as dual in all of them in origin and 
disappearance. Their colligative properties have been appealed 
to in order to confirm this observation, but with conflicting 
results, sometimes confirmatory of the chemical evidence, some- 
times contradictory of it, and sometimes too complex for 
confident chemical interpretation. 
I refer here more especially to Avogadro’s proposition, which 
is in effect that equal volumes of gases are chemically equal or 
molecular. As in the case of Dalton’s atomic theory, there is 
to be distinguished in this proposition what Avogadro really put 
forward as new from what he took for granted. Admitting, as 
was to him a matter of course, that gases have in equal volumes 
equal numbers of particles, he asserted that in the case of 
elementary substances these particles are not the atomic particles, 
but, as in the case of compound substances, particles compounded 
of these, which interact with the particles of other gases as 
chemically equal each to each. If now this proposition is 
divested of all hypothesis, all reference to the mechanical 
structure of gases, it becomes the law that equal volumes of 
gases at the same temperature and pressure, whether simple or 
compound, are almost exactly chemically equal quantities, and 
once in possession of this law we find nothing becomes clearer 
by assuming that equal volumes of different gases contain the 
same number of chemically equal particles. This law is, 
obviously, an advance upon Gay-Lussac’s law similar to that 
of the chemical molecular theory upon the atomic theory of 
Dalton. Unfortunately, however, it does not hold good in the 
case of not a few simple substances, and it seems impossible 
from the chemical point of view, and consistently with the 
molecular theory, to admit that, because the gas-volume has 
only half the expected mass, the chemical molecule of sodium 
or mercury is not bipartite like that of hydrogen or oxygen, and 
chemically equal to either. 
The dual constitution or chemically compound nature of the 
simple substances as thus established by the part they take in 
chemical interactions furnishes further evidence of the unten- 
ability of the belief that the molecule is chemically composed 
of two substances, or their substitutes, simpler than itself, when 
we consider that, were this true, there would be chemical 
union between two things perfectly alike, two portions of the 
same thing. This difficulty was, I believe, first raised by 
Berzelius, and has never been met. Physically, the matter is 
simple enough, if motion in the opposite direction is not counted 
as a difference between two masses. But this would be a non- 
elective union, whilst chemical union is elective. 
The difficulty, insurmountable when made, does not arise 
when the fact is recognised that every chemically single sub- 
stance, whether simple or compound, is, as a substance, one 
and without parts, and can never, therefore, be built up of or 
broken down into parts different from itself. One substance 
(as two molecules) or two substances change into two others or 
into two molecules of one, in an interaction which is instant, 
uninterrupted, and irresolvable into stages, where the inter- 
action is single in character. But just as a body can be 
mentally analysed (as in the investigations of dynamics) into 
mass and motion, which apart are unknown, and as these again 
can each be conceived of as further divided, resolved, condensed, 
and otherwise qualified as centres of mass, compounded motions, 
and so forth, so the chemist is enabled mentally to find 
quantitatively defined this, that, and the other mark of the 
many chemical interactions which have or may have gone to bring 
it into existence, and will or may again have place in the possible 
forms of its dissolution into others. The two methyls in the con- 
stitution of ethane, about which we are quite certain, are not 
two things held together till some interaction sunders them in 
